Combining Dynamic Clustering and Scheduling for Coordinated Multi-Point Transmission in LTE

A promising concept to increase the efficiency of future cellular networks is to facilitate cooperation between Base Stations (BSs), which is known as Coordinated Multi-Point (CoMP). Various coordination techniques have been proposed and are partly standardized in Long Term Evolution (LTE). In this paper we focus on downlink Joint Transmission (JT), because this scheme offers the highest increase of spectral efficiency. In JT the same data is transmitted from multiple BSs such that the signals interfere constructively at the receivers. A major drawback of CoMP is the increased effort for channel measurement and precoding before data transmission. One way to cope with the additional complexity of CoMP is to group BSs into CoMP clusters. Cooperation of BSs is then only allowed between BSs belonging to the same cluster. In this paper we propose a new dynamic clustering algorithm. The novelty of the algorithm is that it not only defines clusters according to the current situation in the network, but also provides hints to the schedulers how to best serve the mobiles. Thereby the complexity of the scheduling process is reduced. The approach is evaluated in a realistic urban scenario, including a traffic model of Web-traffic and vehicular user mobility. The performance of the proposed clustering algorithm is compared with the performance of a traditional network without any cooperation as well as with two static clustering variants.